Analytical and numerical investigation of transient behaviour in hydraulically fractured tight gas reservoirs

Abstract

Development of tight gas reservoirs has long been affected by a lack of understanding of the complex flow profiles, and of the impact of very low permeability on reservoir productivity. Most well tests of tight dry gas wells are of short duration; well test interpretation often becomes difficult and non-unique when the duration of pressure buildup is too short. Thus, such tests typically lead to multiple interpretations and open-ended conclusions. The issue becomes even more complex as simulation software fails to adequately model flow in fracture networks. Presented and investigated here are a series of well tests conducted in tight, naturally fractured, dry gas sandstone reservoirs for long durations: 100 hours in most cases, and 1,000 hours in two extended well tests. Well test responses from these tests were ambiguous. Surprisingly, the transient pressure analysis signature of these tight, naturally fractured, dry gas reservoirs is similar to the signature from lean gas condensate reservoirs. Lean gas condensate reservoirs usually exhibit a two- or three-mobility radial composite model response in the buildup analysis because of the existence of a condensate bank. However, this cannot be the explanation in the dry gas example. Possible causes for the similarity of well test analysis signatures of post-fractured, tight, dry gas and lean gas condensate reservoirs are discussed. The primary objective is to document and characterize the observed behaviour of well tests in unconventional (tight) fields and, secondly, to determine how to analyse these test data, with the goal of obtaining the parameters of the induced fracture and the discrete fracture network (DFN). Primary investigation has explored the possible direct causes (i.e., those related to wellbore, geology, or specific well conditions) that could result in the ambiguous well test response. However, because the well test response has been found repeatedly in a variety of regions around the world, it may not be related to any of those factors. This case study was approached as a direct problem by making a series of conceptual assumptions and investigating their impact on the well test response. Within this context, the possible impact of scale-dependency of properties of fracture networks on the well test response is considered. It is argued that the anomalies in the well test response may reflect a step in the scale-dependent properties of the fracture network. Results suggest that the response reflects the scale-dependence of the intrinsic permeability of the fracture network and a geomechanical effect due to the induced hydraulic fractures. This scale-dependent property has its reflection typically in a change in permeability with relation to distance and time of flow and thus should be considered whenever planning to induce a hydraulic fracture in a natural fracture network.